Protein evolution is a known technology for selection and directed evolution of proteins from large libraries. The basic principle of selection is to ensure that there is a linkage between a specific phenotype (protein) and its encoding, genotype. This phenotype-genotype linkage can be realized in three different ways:                covalent linkage such as mRNA display, and to some extent phage display, bacterial display, yeast display etc.,        non-covalent linkage which use affinity interaction. Examples are ribosome display, CIS display, plasmid display etc.,        compartmentalization such as in vitro compartmentalisation (IVC), compartmentalized self-replication (CSR), simple bacterial screening, high throughput screening etc.        
As indicated in the above paragraph, one example of covalent phenotype-genotype linkage is achieved using mRNA display. As described by Roberts and Szostak (1997) covalent fusions between an mRNA and the peptide or protein that it encodes can be generated by in vitro translation of synthetic mRNAs that carry puromycin, a peptidyl acceptor antibiotic, at their 3′ end.
Non-covalent linkage between phenotype and genotype can be achieved with ribosome display. In ribosome display, an array of RNAs including one or more encoding a target protein is subjected to an in vitro translation system so as to form a non-covalent ternary complex of ribosome, mRNA and protein coupled to tRNA. This array of ternary complexes must be stabilized. Accordingly, each ternary complex formed during in vitro translation uses mRNA lacking a STOP codon at the end. The ternary complexes are further stabilized at low temperature (4° C.) and high concentration of magnesium (50 mM). In the stable complex the linkage between phenotype and genotype is preserved. A selection step follows whereby target protein is selected on the basis of a property of the protein whilst still attached to the ternary complex. Selected ternary complexes may then be disassembled and the mRNA associated with the target protein is amplified by RT-PCR.
In general, ribosome display is successfully applied for the selection of peptides (Mattheakis et al., 1994; Matsuura and Pluckthun, 2003) and proteins (Hanes and Pluckthun, 1997; He and Taussig, 1997; Irving et al., 2001), which bind to different targets. In some cases it is possible to use ribosome display to select for enzymatic activities, performing affinity selection of proteins with suicide inhibitor (Amstutz et al., 2002) or active site ligand (Takahashi et al., 2002).
In in vitro compartmentalization (IVC) phenotype and genotype linkage is realized by in vitro compartmentalization of genes in a water-in-oil emulsion. The number of genes used to prepare the emulsion is calculated such that most of water compartments contain no more than a single gene. The compartmentalized genes are transcribed and translated. The activity of the synthesized proteins is then assessed. Subsequent selection on the basis of protein activity results in amplification of DNA encoding active proteins with desired properties. Water droplets used in most IVC applications are 2-3 μm in size, giving ˜5 femtoliters reaction volume and ˜1010 water-in-oil compartments (50 μl water phase) per 1 ml of emulsion. The first successful example of IVC selection system was based on target-specific DNA methylation activity (Tawfik and Griffiths, 1998). Genes of Haelll methyltransferase were compartmentalized, transcribed and translated. In vitro synthesized methyltransferase in the presence of a cofactor was able to methylate its own DNA. Based on methylated DNA (reaction product) resistance to digestion with Haelll restriction endonuclease, genes encoding methyltransferase were selected from a 107 fold excess of other DNA molecules.
To date many more modifications of IVC have been designed and realized. The easiest way to perform IVC selection is to use DNA-modifying enzymes, in particular DNA-methyltransferases (Lee et al., 2002; Cohen et al., 2004). A similar experimental strategy was applied in order to select for active variants of FokI restriction endonuclease (Doi et al., 2004). Compartmentalized DNA, encoding active restriction endonucleases, was digested and selected by subsequent incorporation of biotin-dUTP and binding to streptavidin beads.
A different IVC selection strategy was applied for evolution of DNA polymerases (Ghadessy et al., 2001; Ghadessy et al., 2004; Ong et al., 2006). This new selection method is based on ‘compartmentalized self-replication’ (CSR) of genes encoding active DNA polymerase. Contrary to usual IVC, where proteins of interest are expressed in situ, CSR is performed by compartmentalization of bacterial cells expressing thermophilic DNA polymerase. Cells resuspended in PCR buffer, supplemented with primers and dNTP's, are emulsified yielding compartments ˜15 μm in size. Each water droplet serves as a separate PCR compartment. During the initial PCR denaturation step the bacterial cells are broken releasing the expressed thermophilic DNA polymerase and its encoding gene into the reaction mixture, allowing self-replication to proceed meanwhile other bacterial proteins are denatured by high temperature.
A modification of IVC is the use of double water-in-oil-in-water emulsions. Water droplets surrounded by an oil layer can be analyzed and sorted in a FACS at a rate of >104 variants per second (Bernath et al., 2004; Mastrobattista et al., 2005).
Selection of proteins for binding may also be performed by IVC. A protein of interest expressed in water-in-oil compartments is coupled covalently (Bertschinger and Neri, 2004) or non-covalently (Doi and Yanagawa, 1999; Yonezawa et al., 2003; Sepp and Choo, 2005) to the gene that encodes it.
Other applications of IVC are known where microbeads entrapped into compartments are used as intermediators for protein and gene coupling. Single genes attached to microbeads are transcribed and translated in water droplets. Newly synthesized protein is captured onto the same bead within reaction compartment. After emulsion is broken isolated beads can be used further for affinity selection (Sepp et al., 2002). Emulsions usually are broken using organic solvents (hexane, ether, chloroform), which can also decrease activity of certain enzymes displayed on the beads, limiting the application of this technology. For selection of catalytic activity microbeads can be easily washed, resuspended in different reaction buffer and compartmentalized again by second emulsification step (Griffiths and Tawfik, 2003). However sometimes rigid enzyme-bead-gene complexes (because of sterical hindrance and enzyme mobility limitations) can not fulfil essential reaction requirements and the activity of the attached enzymes may be lower than that of the free enzyme. In addition, such methods are technically complicated since many additional components have to be used (i.e. affinity tags, antibodies, and beads).
The composition of emulsions used in IVC is designed to ensure stability of water compartments and efficient in vitro transcription of mRNA and subsequent translation of target proteins. In vitro evolution has a broad range of targets to be improved. Some proteins and enzymes of interest are robust hard-workers and can be active enough in different buffers, in particular in reaction mixtures used in IVC. However there are many complicated enzymes, which can work only under optimized or specific conditions. In addition to that the first law of in vitro evolution says—“you will evolve what you are selecting for”. That means, an enzyme evolved and optimized in transcription/translation reaction mixture will work well in that particular mixture and most likely will perform much worse in its own buffer. In some cases enzyme working conditions are incompatible with in vitro transcription and translation mixture used for protein expression in compartments. A partial solution is a nano-droplets delivery system (Bernath et al., 2005) used to transport different solutes into emulsion compartments. Even more sophisticated manipulations with water-in-oil compartments can be done employing microfluidics devices. Highly monodisperse single or double emulsions (Thorsen et al., 2001; Okushima et al., 2004) can be prepared at a rate of up to 10000 aqueous droplets per second. Generated water compartments can be transported in microfluidic channels, fused, subdivided and sorted (Song et al., 2003; Link et al., 2006). Nevertheless full buffer exchange in the compartments is still a problem.
Reverse transcriptases are very important commercial enzymes used to synthesize cDNA from an mRNA target. A lot of research has been done in order to improve properties of reverse transcriptases. However no properly working selection system suitable for in vitro evolution of reverse transcriptase was known to date. Almost all improvements are made and mutants of reverse transcriptase selected using high throughput screening and rational design.
Neither ribosome display (RD), nor in vitro compartmentalization (IVC) may be used for selection of fully active reverse transcriptase. Ternary complexes used in ribosome display usually are not stable at higher temperatures required to perform reverse transcriptase selection and as a consequence linkage between phenotype and genotype will be lost. Whilst relatively stable ternary complexes used in ribosome display can be produced using synthetic in vitro translation extract WakoPURE (Matsuura et al., 2007), in vitro translated reverse transcriptase immobilized to ribosome and mRNA will encounter significant steric hindrance during synthesis of full-length cDNA. There is also the possibility that immobilized enzyme will act in trans as well as in cis, which again is incompatible with protein evolution strategy, because the phenotype-genotype linkage will not be preserved.
IVC generally employs DNA as genetic material. In vitro transcription of mRNA and target protein translation is performed in spatially separated emulsified water compartments in the presence of the DNA. In the case of reverse transcriptase in vitro evolution, the presence of coding DNA sequence abolishes the main prerequisite of selection for reverse transcriptase activity—cDNA has to be synthesized de novo. In other words, selection for better reverse transcriptase variants is based on enzyme ability to synthesize their own coding cDNA from mRNA. Newly synthesized cDNA has to be amplified by PCR, therefore cDNA should be the only source of DNA in the reaction. DNA used in IVC selection will amplify together with cDNA canceling basic selection scheme.
More sophisticated variants of in vitro compartmentalization, such as usage of microbeads entrapped in water compartments (Sepp et al., 2002; Griffiths and Tawfik, 2003), allow complete exchange of reaction buffer. In this approach the phenotype-genotype linkage is realized via microbeads yielding a rigid selection unit mRNA-microbead-protein, which, in case of reverse transcriptase selection, again can cause steric hindrance and as a result inefficient cDNA synthesis.
Taq DNA polymerase able to synthesize ˜300 nucleotides long cDNA was selected using modification of phage display technology (Vichier-Guerre et. al, 2006). Although this approach works, it has some shortcomings: 1) not all proteins can be displayed on phages; 2) absolute requirement to use biotin labeled nucleotides for selection; 3) displayed enzyme can work in trans as well as in cis; 4) because of steric hindrance and enzyme mobility limitations phage-enzyme-DNA/RNA complex can interfere with efficient synthesis of cDNA.
The possibility of selection for reverse transcriptase is also mentioned in WO0222869, which is related to the compartmentalized self-replication (CSR) method. CSR technology is used to select thermophilic DNA polymerases, in particular Taq DNA polymerase (Ghadessy et al., 2001; Ghadessy et al., 2004; Ong et al., 2006). Bacterial cells, expressing mutants library of thermophilic DNA polymerase are suspended in PCR mixture and emulsified yielding separate PCR compartments for in vitro selection of more active polymerases.
Real selection for reverse transcriptase activity will be prevented by the presence of bacterial RNases, which remain active at moderate temperatures and will degrade target mRNA. There will also be DNA contamination from non-selected plasmid DNA (released from bacterial cell) as well as the presence of all E. coli enzymes, structural proteins, ribosomes, NTP, RNases, DNases and small molecular weight molecules.